Abstract Implantable cardioverter defibrillator (ICD) for primary prevention of mortality is an approved therapy in nonischemic cardiomyopathy (NICM) patients with ejection fraction (EF) ?35%. However, the DANISH ICD trial enrolled 1,116 patients with NICM and EF ?35% randomized to ICD (n=556) or no ICD (n=560), demonstrated lack of significant mortality reduction in ICD patients when compared to non-ICD patients (hazard ratio=0.87; p=0.28), but significant reduction in the risk of sudden cardiac death (hazard ratio=0.50; p=0.005). Data from prior clinical trials leading to ICD indications in NICM showed a nonsignificant but clinically meaningful reduction in mortality, and significant reductions in sudden cardiac death (DEFINITE: 458 patients; HR=0.65; p=0.08 and a subgroup analysis of the SCD-HeFT: 792 patients; HR=0.73; p=0.06). The DANISH patients were on optimal pharmacological therapy and 58% had CRT devices. Their 3-year mortality in non-ICD arm was around 10% vs. around 20% in DEFINITE and SCD-HeFT trials, which were conducted 10 years earlier, a reflection of different patient populations and different underlying therapies. Since arrhythmogenic myocardial substrate is characterized by ventricular arrhythmias detected in long-term ECG recordings, we analyzed data from MADIT-CRT trial in 416 NICM patients in NYHA class II with QRS ?130 ms who were randomized to CRT-D and who had a 24- hour Holter ECG recording prior to device implantation. NSVT on preimplantation Holter (observed in 194 [47%] patients) was associated with hazard ratio of 3.08 (p<0.001), enlarged LVESVi >86 ml/m2 was associated with hazard ratio of 2.81 (p<0.001), and Non-LBBB QRS morphology (observed in 44 (11%) patients) with hazard ratio of 3.33 (p<0.001) in a multivariate model predicting fast VT?188 bpm or VF. This risk prediction model could guide therapeutic decisions regarding implantation of CRT-D, CRT-P, or ICD devices in NICM heart failure patients with QRS?120 ms. In specific aim 1 of this proposal, we propose to validated this MADIT-CRT risk stratification model in 400 NICM with wide QRS?120ms. With an increasing interest in observations that late gadolinium enhancement assessed in cardiac MRI is associated with increased risk of sudden cardiac death in NICM patients, we also aim to: 2) determine whether abnormal cardiac magnetic resonance (CMR) imaging with late gadolinium enhancement will be predictive for fast VT/VF and will further improve risk stratification in NICM with QRS?120 ms, and 3) to evaluate cost- effectiveness of the proposed risk stratification approach leading to utilization of less costly devices (CRTP vs. CERT-D).